est_msaeOBns: EBLUPs Optimum Benchmarking for Non Sampled Area based on a...

In yas-q/msaeOB: Optimum Benchmarking for Multivariate Small Area Estimation

EBLUPs Optimum Benchmarking for Non Sampled Area based on a Multivariate Fay Herriot (Model 1)

Description

This function gives EBLUPs optimum benchmarking for non sampled area based on multivariate Fay-Herriot (Model 1)

Usage

est_msaeOBns(
  formula,
  vardir,
  weight,
  cluster,
  samevar = FALSE,
  MAXITER = 100,
  PRECISION = 1e-04,
  data
)

Arguments

formula	an object of class list of formula describe the fitted models
vardir	matrix containing sampling variances of direct estimators. The order is: var1, cov12, ..., cov1r, var2, cov23, ..., cov2r, ..., cov(r-1)(r), var(r)
weight	matrix containing proportion of units in small areas. The order is: w1, w2, ..., w(r)
cluster	matrix containing cluster of auxiliary variables. The order is: c1, c2, ..., c(r)
samevar	logical. If TRUE, the varians is same. Default is FALSE
MAXITER	maximum number of iterations for Fisher-scoring. Default is 100
PRECISION	coverage tolerance limit for the Fisher Scoring algorithm. Default value is 1e-4
data	dataframe containing the variables named in formula, vardir, and weight

Value

This function returns a list with following objects:

eblup

a list containing a value of estimators

• est.eblup : a dataframe containing EBLUP estimators

• est.eblupOB : a dataframe containing optimum benchmark estimators

fit	a list containing following objects:

• method : fitting method, named "REML"

• convergence : logical value of convergence of Fisher Scoring

• iterations : number of iterations of Fisher Scoring algorithm

• estcoef : a data frame containing estimated model coefficients (beta, std. error, t value, p-value)

• refvar : estimated random effect variance

random.effect	a data frame containing values of random effect estimators
agregation	a data frame containing agregation of direct, EBLUP, and optimum benchmark estimation

Examples

## load dataset
data(datamsaeOBns)

# Compute EBLUP & Optimum Benchmark using auxiliary variables X1 and X2 for each dependent variable

## Using parameter 'data'
## Not run: 
Fo = list(f1 = Y1 ~ X1 + X2,
          f2 = Y2 ~ X1 + X2,
          f3 = Y3 ~ X1 + X2)
vardir = c("v1", "v12", "v13", "v2", "v23", "v3")
weight = c("w1", "w2", "w3")
cluster = c("c1", "c2", "c3")

est_msae = est_msaeOBns(Fo, vardir, weight, cluster, data = datamsaeOBns)

## Without parameter 'data'
Fo = list(f1 = datamsaeOBns$Y1 ~ datamsaeOBns$X1 + datamsaeOBns$X2,
          f2 = datamsaeOBns$Y2 ~ datamsaeOBns$X1 + datamsaeOBns$X2,
          f3 = datamsaeOBns$Y3 ~ datamsaeOBns$X1 + datamsaeOBns$X2)
vardir = datamsaeOBns[, c("v1", "v12", "v13", "v2", "v23", "v3")]
weight = datamsaeOBns[, c("w1", "w2", "w3")]
cluster = datamsaeOBns[, c("c1", "c2", "c3")]

est_msae = est_msaeOBns(Fo, vardir, weight, cluster)

## Return
est_msae$eblup$est.eblupOB # to see the Optimum Benchmark estimators

## End(Not run)

yas-q/msaeOB documentation built on June 23, 2022, 7:10 p.m.